The query tree protocol and AQS may be in?¬‚uenced by the distribution of IDs because
they use tag IDs for splitting a tag set. To quantify the similarity of IDs, we de?¬?ne an
identical bit as the length of the identical pre?¬?x all tag IDs have. The tag ID is depicted by
x1x2 . . . xaxa??1 . . . x96 (xi is a binary digit, 1 a<96) and all tag IDs have the same
x1x2 . . . xa if the identical bit is a and each tag has a 96 bit ID. Figure 8.9 gives the
simulation results for various identical bits from 0 (IDs are completely randomly selected)
to 80. We normalize the measured values by the number of recognized tags. As the
identical bit increases, the query tree protocol rapidly degenerates as expected. The
query tree protocol has the highest communication overhead because the reader transmits
all queries causing collisions in every frame. On the other hand, the performance of AQS is
not seriously affected by the similarity of IDs. Since candidate queue CQ excludes queries
of collision cycles of the last frame, AQS uses a collision query only once. However, as the
identical bit increases, idle cycles in the trees of the query tree protocol increase. When
the identical bit is >48, AQS has longer delay than the binary tree protocol because of an
increment of idle cycles. ABS and the binary tree protocol are not affected by the identical
bit because they do not use the patterns of IDs.
Pages:
287
288
289
290
291
292
293
294
295
296
297
298
299
300
301
302
303
304
305
306
307
308
309
310
311